Table 2: Comparison of nonimaging modalities in the detection of intracranial stenosis.
ModalityDegree of stenosisSensitivitySpecificityLimitations
Digital subtraction angiographyInvasive test: Procedure risk rate: 0.3% for all complications, 0.03% for stroke [56]
MRA (TOF) [54] for ICA disease50%–69%37.9%92.1%Limited spatial resolution, flow signal intensity loss as a result of saturation or phase dispersion, susceptibility artifacts near sphenoid sinus, and over- and underestimation of stenosis due to dephasing artifacts
>70%–99%91.2%88.3%
ICA occlusion94.5%99.3%
MRA (TOF) 3T [57]50%–99% stenosis78%–85%95%
Occlusion100%99%
MRA (CE) [54]50%–69%65.9%93.5%
>70%–99%94.6%91.9%
ICA occlusion99.4%99.6%
CTA* [59]Stenosis‡98%99%
Occlusion100%100%
Transcranial Doppler Ultrasound [62]>50% stenosis or occlusion High level of technical and procedural skill is required to obtain the best quality images. Reliable insonation of the posterior circulation is particularly difficult
For MCA stem (M1)90%–99%90%–99%
For intracranial segment (V4) of vertebral and basilar artery70%–80%90%–99%
CDDI [61]Atheromatous pseudo-occlusion
Unenhanced†70%92%False negative rate 30%
Echo-enhanced PFI83%92%False negative rate 17%
Unenhanced95%92%False negative rate 5%
Echo-enhanced94%100%False negative rate 6%
CDDI: Color Doppler-assisted duplex imaging, PFI: power-flow imaging.
*Data are percentages using DSA as the reference standard.
‡North American Symptomatic Carotid Endarterectomy Trial (NASCET) criteria were used for stenosis calculations: [ ( 𝐷 𝑛 − 𝐷 𝑠 ) / 𝐷 𝑛 ] × 1 0 0 , where 𝐷 𝑛 is normal diameter and 𝐷 𝑠 is stenosed diameter. NASCET stenoses were grouped according to the following grading scale: normal (0%–9%), mild (10%–29%), moderate (30%–69%), severe (70%–99%), or occluded (no flow detected). Normal (0%–9%) and mild (10%–29%) stenosis were not considered diseased vessel segments and were excluded from analysis.
†Ultrasound emission energy and gain cannot be increased high enough without the appearance of disturbing acoustic noise that diminishes the reliable depiction of orthograde flow signals.